Biological Energy Systems

Established Biological energy systems is the proposal that living cells can be a working component of an energy supply, and after 115 years of trying the measured record is unusually easy to state. M.C. Potter drew current from Saccharomyces cerevisiae in 1911. Established Barnett Cohen stacked microbial half-cells in 1931 and reached over 35 V at 2 mA — roughly seventy milliwatts, which is to say that the power problem was fully visible in the first two decades and has not been solved since. Established The best areal power density reported for a microbial fuel cell today is 0.83 W/m² for conventional two-dimensional electrodes and 7.72 W/m² for the best miniaturised device, against roughly 200 W/m² at peak for silicon photovoltaics on the same footprint.

Established The number the field actually quotes is 11,220 W/m³, and it is the wrong metric. That record, set with a three-dimensional graphene macroporous scaffold anode, is volumetric, and it was obtained by miniaturisation — shrinking the reactor improves the volume denominator while doing nothing whatever for the electrode area in the numerator. Established Capital cost in an electrochemical reactor scales with electrode and membrane area, not with volume, so the volumetric record is close to economically meaningless. Frontier The areal series tells the real story: nanowatts per square metre in 2002, about one watt per square metre by 2011, and essentially flat since. A brief that reports the volumetric figure alone has inverted its subject's status.

Established The second thing a reader needs is that the founding structural model of this field was overturned. For roughly two decades Geobacter's conductive filaments were held to be Type IV pili, and a large electrode-engineering literature reasoned from that. Established Cryo-electron microscopy established that the pili are secretion pseudopili and the conductors are cytochrome filaments — and that the pili are approximately 20,000-fold less conductive than OmcZ. Speculative The framing that organises everything below is that biology loses by two to three orders of magnitude whenever the product is electrons, and that the only defensible question is whether it can win when the product is a molecule.

Established Start with the historical performance table, because it contains the field's central embarrassment and almost nothing else needs to be argued once it is on the page. Miniaturised microbial fuel cell performance runs: Chiao and colleagues at 5.95 nW/m² in 2002; Siu and Chiao at 4 mW/m² and 40 W/m³ in 2008; Qian and colleagues at 1.5 mW/m² and 15 W/m³ in 2009; Parra and Lin at 0.12 W/m² the same year; Choi and colleagues at 47 mW/m² and 2,333 W/m³ in 2011; Ren and colleagues at 11,220 W/m³ in 2016; Jiang and colleagues at 0.894 W/m² and 745 W/m³ in 2017; Gao and colleagues at 22.12 W/m³ in 2020. Established Six orders of magnitude of areal improvement between 2002 and 2011, and then a plateau; everything spectacular in the table after 2011 is volumetric. Established The two records that matter are stated plainly in the same review: 0.83 W/m² and 3,320 W/m³ for two-dimensional electrodes, 7.72 W/m² for the miniaturised class.

Established The areal-versus-volumetric distinction is not a quibble about units; it is the difference between a technology and a laboratory curiosity. A microbial fuel cell is a reactor with an anode, a cathode, usually a membrane, a substrate feed and a biofilm that has to be kept alive, and every one of those cost items scales with area or with throughput. Speculative For the volumetric record to translate into economics there would have to be a manufacturing route in which reactor volume rather than electrode area dominates cost, and none is known. Established Against this, the comparison that the field mostly declines to run: photovoltaic modules convert about twenty per cent of incident light and need no feed, no membrane, no sterility and no biology. Speculative Wherever the product is electrons, the microbial route is behind by two to three orders of magnitude on the axis that governs land and capital.

Established Scale-up is where the arithmetic becomes vivid. A 10 litre brewery-wastewater prototype ran at the University of Queensland in 2007; sediment microbial fuel cell tests exceeded 150 litres by 2015; the current published scale-up frontier is a 20 litre stack with a low-cost membrane, coupled to a microbial electrolysis cell for hydrogen, reported in 2026. Established After 115 years the field's reactors are the size of a domestic water heater. Frontier The one place where the numbers are genuinely competitive is not power at all: microbial desalination has been reported at 0.5 kWh/m³, an eighty-five per cent reduction against conventional processes, and microbial-fuel-cell biosensing is now a review-level subfield in its own right. Speculative The honest reading is that this is good electrochemistry attached to the wrong business case, and that its commercial future, if it has one, lies where the electricity is a by-product of a service someone already pays for.

Established Underneath the devices sits the actual science, which is extracellular electron transfer, and this is where the field's most instructive correction lives. Geobacter sulfurreducens holds one of the highest recorded in vitro current densities, and biofilm conductivity correlates positively with current density — the observation that launched two decades of anode design. Established The named constraints on further gains are low absolute current against industrial sources, low intra-biofilm pH degrading electron transfer, and non-conductive appendages interfering with it. Established For years the conductive filaments were attributed to Type IV pili, until Gu and colleagues, in Nature in 2021, established that G. sulfurreducens pili are PilA-N/PilA-C heterodimers functioning as secretion pseudopili, and that the electron-conducting structures are cytochrome filaments of OmcS and OmcZ. Established The quantitative statement is the one to keep: the PilA filaments are approximately 20,000-fold less conductive than OmcZ nanowires, and they lack the aromatic stacking that efficient electron transport requires. Speculative The pili do not merely fail to be the nanowire; they are four orders of magnitude away from being it, and a substantial body of electrode-engineering rationale was built on the assumption that they conduct at all.

Established The most remarkable biological electron conductor known is not in a fuel cell and was not engineered. Cable bacteria, discovered in 2010 when Nielsen, Risgaard-Petersen and colleagues found sulfide oxidation and oxygen reduction spatially separated in marine sediment and coupled faster than chemical diffusion could explain, conduct electrons over centimetres — roughly ten thousand cell lengths. Established The filaments are one to four microns in diameter and over a centimetre long. Established Fifteen to sixty electron-conducting fibres of about fifty nanometres run lengthwise inside a shared outer membrane, built from proteins rich in nickel and sulfur. Established Within a fifteen-centimetre sediment layer, cable bacteria can supply up to two kilometres of total filament length per square centimetre. Frontier Two confirmatory experiments carry the mechanism: Pfeffer and colleagues in 2012 demonstrated conduction across non-conductive glass microspheres, which excludes a diffusible shuttle, and Bjerg and colleagues in 2018 measured single-filament conductivity by Raman microscopy. Speculative Nothing in the applied literature has yet turned this into a device, which matters because it is the one place in the subject where biology does something no engineered material does.

Established Running the cell backwards — microbial electrosynthesis — is the strongest idea in the subject, and its founding result is sixteen years old. Nevin, Woodard, Franks, Summers and Lovley reported in mBio in 2010 that the acetogen Sporomusa ovata formed a cathode biofilm which consumed electrons to produce acetate and small quantities of 2-oxobutyrate, accounting for over eighty-five per cent of the electrons consumed. Established The paper explicitly framed the approach as an innovative form of artificial photosynthesis when coupled to photovoltaics. Established Read the number carefully: eighty-five per cent is a coulombic efficiency — the fraction of delivered electrons ending up in product — not an energy efficiency and not a cost. Established The product was acetate, a commodity in the several-hundred-dollars-per-tonne range, from a laboratory cathode. Frontier Sixteen years later there is no commercial microbial electrosynthesis plant. Speculative The pattern is identical to the microbial fuel cell pattern: an excellent normalised laboratory number attached to a metric that does not govern cost.

Established Artificial and bio-hybrid photosynthesis have the longest lineage and the widest gap between laboratory and deployment. Giacomo Ciamician proposed artificial photosynthesis in 1912; Akira Fujishima discovered titanium dioxide photocatalytic water splitting in the late 1960s; William Ayers demonstrated visible-light water splitting with multijunction semiconductors in 1983. Established The highest reported laboratory prototype efficiency is 22.4 per cent. Established Daniel Nocera's 2011 artificial leaf — a playing-card-sized silicon cell with a cobalt-phosphate catalyst on one face and a nickel-based catalyst on the other, splitting water without wires — is the best-known artefact in the area. Frontier Its bio-hybrid successor, the Liu, Colon, Ziesack, Silver and Nocera water-splitting biosynthetic system published in Science in 2016, feeds hydrogen from electrochemical water splitting to bacteria that fix carbon dioxide; its title claims carbon dioxide reduction efficiencies exceeding photosynthesis, and that title has done more work in circulation than the paper has.

Established The denominator for every one of those comparisons is the number readers most often get wrong. Natural photosynthesis has a theoretical maximum of 4.6 per cent solar-to-biomass for C3 plants and 6.0 per cent for C4 plants; real-world plant efficiency is usually below one per cent. Established Practical algal solar conversion runs at about three per cent, with five to seven per cent considered theoretically attainable in a controlled system. Speculative This is the structural argument that biological energy conversion is capped: light capture is not free, the Calvin cycle and RuBisCO are not efficient, and a living cell spends most of the captured photon budget on staying alive rather than on making product. Frontier The bio-hybrid answer — a semiconductor for light capture, where physics gives twenty per cent, and a cell only for carbon fixation, where biology gives selectivity — is the one serious route around the cap, and it converts a photobiology problem into an interface and containment problem.

Established Biofuels are the part of this subject that actually operates at scale, and their arithmetic is the least flattering. Global bioethanol supplies about 2.2 EJ per year. Established Up to forty per cent of the United States corn crop goes to ethanol, and roughly ten per cent of all world grain is turned into biofuel; IEA figures for 2021 put ethanol production at twenty per cent of world sugar supplies and thirteen per cent of corn supplies. Established Cellulosic ethanol, the promised escape from the food-versus-fuel trap, cost about $2.65 per gallon in 2007 estimates, roughly two to three times corn ethanol, and collapsed commercially: most plants were cancelled or closed in the early 2010s as the technical obstacles proved too difficult, and as of 2018 exactly one cellulosic ethanol plant remained operational. Established Algae were the successor promise: 58,700 to 136,900 litres of oil per hectare per year, stated as ten to twenty-three times the next-highest-yielding crop, with some species exceeding sixty per cent of dry weight as oil against soybean's two to three per cent, and harvest cycles of one to ten days rather than a year.

Established The algae economics have never closed, and the reason is documented rather than speculative. The United States Department of Energy's Aquatic Species Program ran from 1978 to 1996, spent $25 million over eighteen years, and screened about 3,000 algal strains. Established Its central finding — which the entire subsequent industry rediscovered at far greater expense — is that rapid growth and high lipid production are mutually exclusive: starve algae for nitrogen and they make oil but stop growing. Established The programme was terminated in 1996 as unable to reach petroleum-competitive pricing, against a sub-$20-per-barrel oil price. Established Unextracted algal oil was projected at $59 to $186 per barrel in the 1990s; IEA 2017 figures put algal biomass production at $0.54 per kilogram in open ponds and $10.20 per kilogram in photobioreactors, with recovered oil at about $2.80 per litre even from the cheapest biomass. Established Water consumption runs at 607 to 1,944 litres per litre of biodiesel, and hydrothermal liquefaction can convert about seventy-nine per cent of algal carbon into fuel. Established ExxonMobil ended its algae biofuel research funding in December 2022, the last major oil company invested in the area.

Frontier The visible frontier of microbial fuel cell research is a materials-first anode literature, and its structure should be described before its contents. Recent claims include three-dimensional fold MXene film on nickel foam for high power density and fast start-up, nanoflower molybdenum disulfide anchored on electrospun carbon nanofibre interpenetrated with reduced graphene oxide, metal-organic-framework-derived iron oxide on carbon cloth, iron oxide nanospheres on reduced graphene oxide, and continuous carbon fibres as anode. Frontier Each is a single group; none has been independently replicated at the reported figure. Frontier A meaningful fraction of the recent high-power-density claims are preprints rather than peer-reviewed papers — the human-urine-driven disordered-porous-carbon work and one of the distillery-wastewater studies are both SSRN postings. Speculative The field is producing a high rate of unreplicated single-lab records in incompatible normalisations, which is exactly the condition under which an apparent trend can be an artefact of reporting choice.

Frontier The most consequential open question after Gu is what an anode designed for cytochrome filaments actually looks like. If the conductors are OmcS and OmcZ rather than pili, then surface chemistries selected to bind or template pili were selected against the wrong target, and the correct design objective is to maximise OmcZ filament density and orientation at the interface. Speculative No one has yet published a systematic anode benchmark stated in those terms. Speculative A stronger version of the same question: whether the twenty-thousand-fold conductivity gap means that most of the reported anode improvements since about 2005 are surface-area effects that were rationalised through a mechanism that was not operating.

Frontier Scale-up and power electronics are the two frontiers nobody photographs. The published state of the art in stacked reactors is twenty litres, integrated with a microbial electrolysis cell so that hydrogen rather than electricity is the exported product. Frontier Extracting usable power from a sub-volt, high-impedance, biologically variable source is a genuine and under-addressed problem with its own review literature; a device that produces milliwatts at 0.3 V is not obviously usable even where the biology works. Speculative It is defensible that power management, not anode materials, has been the binding constraint on microbial fuel cell applications for a decade, and that the field has spent its effort on the more publishable half of the problem.

Frontier Biophotovoltaics — cyanobacteria or algae on an electrode, exporting current derived from light rather than from an organic feed — is the sub-field this brief can characterise structurally but cannot yet quantify. Frontier No primary biophotovoltaic performance figure was verified for this brief, and none is printed here; readers should treat the omission as a declared gap rather than as evidence of absence. Speculative What can be said from arithmetic already on the page is that a biophotovoltaic device inherits both of this subject's ceilings at once. Its light capture is bounded by photosynthesis — a theoretical 4.6 to 6.0 per cent, under one per cent in practice for plants and about three per cent for algae — and its electron export is bounded by the same areal plateau that constrains every other microbial electrode, 0.83 to 7.72 W/m². Speculative Two multiplicative caps on the same square metre is the least promising arrangement in the subject, and it is why the serious version of solar-to-fuel biology splits the two functions rather than asking one organism to perform both.

Frontier The applications frontier has quietly pivoted away from generation. Rice-mill and distillery wastewater studies now tie power output to a paid-for treatment service rather than to electricity sales; microbial-fuel-cell biosensors for environmental monitoring and heavy-metal toxicity have their own critical reviews; the desalination figures are the strongest energy numbers in the subject. Frontier This is a healthy adjustment and it is rarely stated as one, because stating it concedes the generation case.

Frontier Two live areas are under-covered here and the gap should be declared rather than papered over. Direct interspecies electron transfer — electron exchange between different species within an aggregate, which is now thought to matter in anaerobic digesters — and electromethanogenesis, in which a cathode drives methane production, were not surveyed for this brief. Frontier Nor was the microbial electrosynthesis scale-up literature between 2010 and 2026: the founding result is in, the intervening sixteen years are not. Speculative A reader should treat the microbial electrosynthesis section of this brief as anchored on a single well-verified 2010 primary and should not assume the absence of progress that could not be checked.

Established The workback plan below is stated for the hardest target in the subject: a biological system supplying a meaningful fraction of primary energy demand. Established Link one, a standardised benchmark. Three laboratories, one substrate, one external resistance, one temperature, reporting areal and volumetric power density and coulombic efficiency together. This is institutional, not scientific, and it is first because nothing downstream can be believed without it. Speculative Link two, break the areal plateau. Demonstrate sustained areal power density above 10 W/m² on real wastewater for more than a thousand hours, against a current best of 0.83 to 7.72 W/m². This is a scientific unknown and it is the binding link for everything electromicrobiological.

Speculative Link three, beat the two-vessel control. For any bio-hybrid or electrosynthesis claim, demonstrate that the integrated device beats electrolysis-then-fermentation in two separate vessels on dollars per kilogram of product. Frontier Almost no paper in the area runs this control, and it is the one that decides whether integrating the electrode and the organism buys anything at all. Speculative Link four, decouple growth from product. Publish a closed carbon balance for any organism claimed to break the growth-versus-lipid trade-off the Aquatic Species Program identified. Speculative Link five, non-sterile continuous operation. Six months of continuous photobioreactor or electrosynthesis operation without axenic conditions and without productivity collapse. This is the underrated link: the algae industry's repeated failures were more often microbiological contamination and harvesting cost than photosynthetic efficiency.

Established Link six, land and water honesty. Independent audited life-cycle assessment reporting litres of water per litre of fuel — against the documented 607 to 1,944 — and hectares per petajoule. Speculative Link seven, a price signal. A sustained carbon or fuel price that makes $2.80-per-litre algal oil competitive. Established Two links bind, and they are different kinds of blocker. Link two is a scientific unknown; link seven is a market condition. Speculative Conflating them is the characteristic error of the field's advocacy: the biofuel case is waiting on a price, the electromicrobiology case is waiting on physics, and a policy instrument that fixes one does nothing for the other. Speculative A third distinction is worth adding, because it decides how much of the chain is worth building: links one, three and six are cheap and could all be completed within five years by a funder with no stake in the existing records, while links two, four and five are open scientific questions with no scheduled answer. Speculative The cheap links are the ones that would tell the expensive ones apart, which is the argument for doing them first even though none of them produces a record.

Established This brief waits on two results that other subjects on this map produce for their own reasons. Frontier From Genetic Engineering it needs the pathway-engineering capability that any of the exotic proposals here presuppose: an organism with constitutively high flux to a reduced product in a nitrogen-replete cell, or an electroactive chassis with cytochrome filament expression tuned as a design variable rather than accepted as a wild-type property. Speculative Neither exists, and the honest statement is that every “engineered electroactive organism” proposal in section three is a claim on genetic engineering's future rather than on electromicrobiology's present.

Frontier From Synthetic Ecosystems it needs the closed-loop energy accounting that closed-system research has already had to do honestly. Established BIOS-3 is the instructive case: a 315 m³ sealed facility that regenerated atmosphere and water completely and reached ninety-three per cent food regeneration, while drawing about 400 kW from a nearby hydroelectric station, with each phytotron originally lit by twenty cylindrical six-kilowatt xenon lamps. Established That is roughly 120 kW of lamps per phytotron to grow food for a crew of three. Speculative The lesson transfers directly: a biological loop that is closed in matter can be extravagantly open in energy, and any biological energy system proposed for a closed habitat has to report the energy it consumes to stay alive, not only the energy it exports.

Speculative One dependency is not a brief and is not a scientific result: an audited reporting standard. The chain of results in section four cannot even be evaluated until power density is reported in one normalisation. That is a constraint on the field's institutions, and it is discussed in section nine rather than typed as an edge here.

Established Six experiments would move this subject more than any plausible anode paper. Speculative One: the round-robin. Three independent laboratories, a registered protocol, synthetic wastewater of stated composition, a fixed external resistance, thirty degrees Celsius, thirty days, reporting anode-areal, cathode-areal, membrane-areal and volumetric power density plus coulombic efficiency for the same device. Cost is well under a million dollars. Speculative It would very likely erase a large fraction of the claimed progress since 2010, which is precisely why no participant has an incentive to fund it.

Speculative Two: the thousand-hour areal test. Any device claiming a record reports it as sustained areal power density on real, unamended wastewater for a thousand hours with the biofilm undisturbed. Frontier The interesting outcome is not the peak but the decay curve, which the literature rarely publishes. Speculative Three: the two-vessel control. Run electrolysis in one vessel and fermentation of the resulting hydrogen in a second, with the same total energy input, and compare dollars per kilogram of acetate or of whatever the integrated device makes. Speculative If the integrated device does not win, microbial electrosynthesis is a demonstration rather than a process, and the field should say so.

Speculative Four: the carbon balance. Anyone claiming to have broken the growth-versus-lipid trade-off publishes a closed carbon balance — carbon in, carbon to biomass, carbon to lipid, carbon respired — because thermodynamically the carbon budget is conserved and the realistic version of the claim is a shift in the trade-off curve rather than its elimination. Speculative Five: the non-axenic six-month run. A photobioreactor or cathode reactor operated continuously without sterility for six months, reporting productivity weekly and the community composition monthly. Speculative Six: the post-Gu anode series. A systematic comparison of anode surface chemistries scored against measured OmcZ and OmcS filament density at the interface, which is the design objective the cryo-EM result implies and which no one has yet run. Speculative Seven: the biophotovoltaic areal benchmark. A photosynthetic electrode reported as watts per square metre of illuminated area under a stated irradiance, alongside the same organism's solar-to-biomass efficiency in suspension, so that the two ceilings this brief argues are multiplicative can be measured against each other rather than assumed.

Established The engineering constraints in this subject are unusually concrete, and three of them are shared with every other large-scale bioprocess. Established The first is that cost tracks electrode area and membrane area, which is why the volumetric record does not help and why a device that must present a square metre of electrode to produce a watt is not a device anyone will install. Frontier The second is power conditioning: a sub-volt, high-impedance, biologically variable source requires boost conversion and storage whose parasitic losses are comparable to the output, and this has its own review literature precisely because it is not solved.

Established The third is the bioreactor scale ceiling, and the best-documented statement of it comes from an adjacent field. Humbird's scale-up analysis of cultured meat — universally cited as the paper proving that cultured meat cannot work — actually estimates $37 per kilogram of wet cell mass for a fed-batch process, $51 per kilogram for perfusion, and $22 per kilogram with plant protein hydrolysate at $2 per kilogram. Established That is expensive, not impossible. Established Its distinctive technical claim is the one the field does not quote: bioreactors larger than 20 m³ have a lower maximum cell density because of carbon dioxide inhibition, and the only way to circumvent that in a sparged vessel is to sparge harder, possibly to the point of cell death. Speculative If that ceiling is general to dense aerobic culture, it constrains every biological energy proposal that assumes large well-mixed vessels, and it is a physical limit rather than a cost curve — which makes it far more interesting than the media-price argument it is usually reduced to.

Established Two more, specific to biofuels. Harvesting dilute algal biomass from water is the dominant cost item and is not improved by better photosynthesis. Established And contamination control at open-pond scale has no cheap solution: the choice is between a closed photobioreactor at $10.20 per kilogram of biomass and an open pond at $0.54 per kilogram that a competing organism can take over in a week.

Established The adjacent technology that matters most is the one this field is racing: photovoltaics plus electrolysis plus storage. Speculative A field whose key metric has been flat for a decade is competing against a cost curve that fell by roughly ninety per cent over the same period, and losing a race against an exponential is not fixed by improving anodes. Speculative The counter is real and should be stated: biology self-assembles, self-repairs, and works at ambient temperature and pressure on dilute, dirty, distributed feedstocks — exactly the conditions in which engineered systems are most expensive. The photovoltaic comparison is only fair where a grid and a maintenance crew already exist.

Established The best evidence that biology wins on molecules rather than electrons comes from fermentation, not from energy. Recombinant chymosin received GRAS status in 1990, the first genetically engineered food product to do so, and now supplies over ninety per cent of the global rennet market. Speculative That is what a successful biological production technology looks like: a high-value molecule that is hard to make any other way, produced in a stirred tank, with no claim on the energy system whatsoever. Established The scale of what an energy-relevant biological process would have to displace is a different order of magnitude: the Haber process produced 230 million tonnes of anhydrous ammonia per year as of 2018, consuming three to five per cent of world natural gas production and about one to two per cent of world energy supply.

Frontier Two further adjacencies. Controlled-environment agriculture supplies the closest available comparison for “how much energy does it take to run photosynthesis indoors”: a 2024 Finnish life-cycle assessment measured 10 to 95 MJ per kilogram of direct energy for vertical farming against 120 to 133 MJ per kilogram for greenhouse lettuce, with the best case dependent on renewable electricity and waste-heat recovery. Established And wastewater treatment infrastructure is the realistic host for microbial fuel cells, because it is the one setting where a slow, area-hungry, biofilm-dependent reactor is already being paid for doing something else.

Established This subject has an institutional problem that is larger than any of its scientific ones, and it is measurement. Power density is reported in at least four incompatible normalisations — anode area, cathode area, membrane area, reactor volume — with no standard substrate, temperature or external resistance. Speculative A field that permits four denominators permits the author to choose the flattering one, and the historical table in section two shows exactly what that produces: a volumetric record that is quoted everywhere and an areal record that has not moved.

Speculative The proposal nobody is funding is a registered-report benchmark. One substrate, one load, three independent laboratories, both areal and volumetric numbers mandatory, results published whatever they show. It would cost under a million dollars, it is technically trivial, and it would probably erase a large fraction of the claimed progress since 2010. Speculative No participant has an upside: not the laboratories whose records would be re-measured, not the journals that published them, not the funders who paid for them. Speculative This is an institutional blocker rather than a scientific one, and it is the highest-leverage thing an institute with no stake in the existing records could do.

Established The Geobacter episode is the second institutional lesson and it is worth stating carefully, because the wrong lesson is easy to draw. A founding structural model was wrong; a large downstream literature reasoned from it; a cryo-EM study corrected it in 2021 and quantified the error at roughly twenty thousand fold. Speculative The wrong lesson is that the field is unserious. The right one is that a structural claim propagates through an applied literature far faster than anyone re-tests it, because the applied papers cite the mechanism as background rather than as a hypothesis under test — and the correction, when it came, came from a different technique in a different discipline rather than from the applied field noticing its own anomalies. Speculative An institute reading this subject should treat “the mechanism everybody cites in the introduction” as the least-tested proposition in any applied literature, not the most.

Established Third, funding scale. The Aquatic Species Program spent $25 million over eighteen years — about $1.4 million a year, less than a single substantial modern grant — and is routinely cited as proof that algal fuel is impossible. Speculative It was not falsified; it was defunded at an oil price that no longer exists. Established ExxonMobil's December 2022 exit is a better piece of evidence about the economics than the 1996 termination, because it was made at a modern price with modern biology.

Established The largest ethical fact in this subject is that biofuels are already a very large land use and are usually discussed as though they were marginal. Up to forty per cent of the United States corn crop goes to ethanol; roughly ten per cent of all world grain is turned into biofuel; IEA figures for 2021 put ethanol production at twenty per cent of world sugar supplies and thirteen per cent of corn supplies. Speculative Whatever else that is, it is a food-system intervention conducted through an energy policy instrument, and the people who bear its price effects are not the people who chose it.

Established Water is the second. Algal biodiesel consumes 607 to 1,944 litres of water per litre of fuel produced. Speculative A technology whose principal siting advantage is that it can use non-arable land is generally being proposed for arid land, which is where that ratio bites hardest. Speculative An audited water and land accounting should be a condition of public funding rather than an afterthought.

Frontier Third, engineered electroactive organisms raise a containment question that fuel cells do not. An organism engineered for high extracellular electron transfer is, by construction, adapted to respiring solid surfaces in sediment and soil — the environments where its wild relatives already dominate. Speculative That is a different escape profile from a fermentation organism selected for a sterile stirred tank, and the field has not seriously argued the case either way. Speculative Fourth, and less obvious: overstated claims in this area have a real cost, because climate capital is finite and a watt claimed for biology is a watt not deployed elsewhere. Speculative Reporting a volumetric record without the areal one is not a stylistic failing; in a funding environment it is a misallocation mechanism. Speculative The corollary is uncomfortable for an institute that funds exotic work: the most defensible ethical position on this subject is to keep funding the science while refusing to fund it as energy policy, and to say publicly which of those two things a given grant is.

Speculative The structural claim worth taking seriously is that biological energy conversion is permanently capped by photosynthesis's own ceiling. The theoretical maximum is 4.6 per cent for C3 and 6.0 per cent for C4 photosynthesis; real plants deliver under one per cent; practical algae reach about three. Established Silicon photovoltaics deliver around twenty per cent and are still improving. Speculative If that gap is structural rather than contingent — and the reason it exists is that a cell must spend its captured photons on maintenance, repair and reproduction rather than on product — then a technological civilisation will never obtain bulk energy from living systems, and the entire subject resolves into chemistry rather than power.

Speculative The residual case for biology is narrow, defensible, and worth stating precisely. Biology wins where the product must be a molecule with structure that electrochemistry cannot cheaply impose; where the feedstock is dilute, dirty and distributed; where the system must self-assemble and self-repair without a maintenance crew; and where energy is a by-product of a service someone is already paying for. Frontier Wastewater treatment, desalination, remote sensing and closed-habitat life support are all in that set. Speculative Bulk electricity generation is not, and no result in this brief suggests it ever will be.

Handwave The interesting long-horizon version is not about efficiency at all. A civilisation that could design an organism whose entire photon budget went to product, with maintenance externalised to an engineered support system, would not be doing photosynthesis in any sense a biologist would recognise — it would be running a chemical plant with self-replicating catalysts. Handwave Whether that object is closer to biology or to manufacturing is a question about words, but it is the only version of “biology beats photovoltaics on an area basis” that is even coherent, and every step of it is asserted rather than demonstrated.

These horizons track results that would change what can honestly be said about moving energy with biology, not forecasts of deployment.

• 10 yr: Frontier The decade's decisive result is negative or it is nothing: either someone demonstrates sustained areal power density above 10 W/m² on real wastewater for a thousand hours, or the plateau at 0.83 to 7.72 W/m² should be treated as the terminal value and microbial fuel cell funding should move wholesale to treatment, desalination and sensing. Frontier Expect the applications pivot to complete regardless, because the treatment-as-product framing already dominates the recent literature. Speculative Expect at least one post-Gu anode paper stated in terms of cytochrome filament density rather than pili, and expect the twenty-thousand-fold correction to remain largely unabsorbed in the applied literature for most of the decade.
• 25 yr: Speculative Microbial electrosynthesis either clears the two-vessel control on dollars per kilogram for at least one product, in which case it becomes a chemical process technology with a real niche, or it does not, in which case it remains a demonstration and the founding 2010 result stands as its high-water mark. Speculative Bio-hybrid photosynthesis faces the same fork in harsher form: it must beat photovoltaics plus electrolysis plus a conventional fermenter, on land, at scale, in the open, without sterility. Speculative Algal fuel returns only on a sustained price signal, and the biological blocker the Aquatic Species Program identified in the 1980s is unchanged until someone publishes a closed carbon balance showing otherwise.
• 50 yr: Speculative By this horizon either the photosynthetic ceiling has been shown to be a regulatory rather than a thermodynamic limit — which would require engineering a cell whose maintenance cost is a small fraction of its captured photon budget, something no organism does — or biology's role in the energy system is settled permanently as feedstock conversion and waste treatment. Speculative Cable bacteria are the wildcard: centimetre-scale conduction through a self-assembling protein structure at ambient temperature is a capability no engineered material has, and fifty years is long enough for it to become either a materials platform or a curiosity. Handwave Nothing currently distinguishes those two futures.
• 100 / 250+ yr: Handwave At this horizon the honest position is that the interesting object is not a better microbe but a hybrid whose light capture is inorganic, whose catalysis is biological, and whose maintenance is paid for from outside — and that such a thing has no established name because nobody has built one. Handwave It is worth ending on the observation that this subject's central metric has not moved since 2011 and that its founding structural mechanism was wrong for twenty years, which together suggest that the binding constraint on a century-scale forecast here is not physics but the field's willingness to measure the same quantity twice.

• Depends on This brief waits on two results that other subjects on this map produce for their own reasons. From genetic engineering it needs a pathway-engineering capability that does not yet exist: an organism with constitutive high flux to a reduced product in a nutrient-replete cell, or an electroactive chassis in which cytochrome filament expression is a tunable design variable rather than an accepted wild-type property. Every exotic proposal in this brief is a claim on that capability. From synthetic ecosystems it needs the closed-loop energy accounting that closed-system research has already been forced to do honestly, because a biological energy device inside a closed habitat has to report the energy it consumes to stay alive and not only the energy it exports. BIOS-3 drew about 400 kilowatts from a hydroelectric station to keep three people fed inside 315 cubic metres, and that number is the correct frame for every proposal here.
• Enables What this subject supplies to others is a benchmark rather than a component, and the benchmark is easy to state and hard to beat: 0.83 to 7.72 watts per square metre for microbial electricity, about three per cent practical solar-to-biomass for algae, and roughly 200 watts per square metre for the photovoltaic alternative on the same ground. Any proposal that puts a living organism in an energy path now has to clear those three numbers explicitly. It also supplies a methodological standard: report areal and volumetric power density together, run the two-vessel control, and publish a closed carbon balance. No typed enabling edge is recorded because a standard is not a technology, and because the subjects that would consume this one's results — wastewater infrastructure, environmental sensing, closed-habitat life support — are consuming a service rather than a component. The one contribution that behaves like a technology is the cytochrome-filament result itself: an electron-conducting protein structure that self-assembles at ambient temperature is a materials fact before it is an energy fact, and it belongs to anyone building bioelectronic interfaces rather than to this brief.
• Adjacent The adjacent subjects share this one's failure modes rather than its components. Future agriculture faces the same photosynthetic ceiling from the other side, and the same contamination and harvesting economics that killed algal fuel. Synthetic biology and designer organisms own the chassis engineering that every exotic proposal here presupposes. Microbiome engineering shares the specific hazard on display in the Geobacter correction, which is an applied literature reasoning confidently from a mechanism that nobody in that literature is re-testing. And photovoltaics with electrolysis and storage is not adjacent at all but directly competing, which is the fact this subject most needs to keep in view.

Established “Microbial fuel cells have reached over 11,000 W/m³, so scale-up is close.” Established That record is volumetric and was obtained by miniaturisation, which improves the volume denominator without improving the electrode area in the numerator, while areal power density, which governs cost, is 0.83 W/m² for two-dimensional electrodes and 7.72 W/m² at the miniaturised record, and it has been roughly flat for a decade. Speculative Quoting the volumetric figure alone does not merely overstate the field's progress; it inverts the field's actual status, presenting a plateau as a breakthrough.

Established “Geobacter nanowires are Type IV pili.” Established They are not. Gu and colleagues showed in Nature in 2021 that G. sulfurreducens pili are PilA-N/PilA-C heterodimers functioning as secretion pseudopili, that the electron-conducting structures are OmcS and OmcZ cytochrome filaments, and that the pili are approximately 20,000-fold less conductive than OmcZ and lack the aromatic stacking efficient transport requires. Speculative What this should teach a reader is not that the field is careless but something more specific and more useful: a founding structural model can support a large applied literature for two decades without being re-tested, because applied papers cite a mechanism as settled background rather than as a hypothesis, and the incentive to check it belongs to nobody. Speculative The correction arrived from cryo-electron microscopy — a different technique, a different community — rather than from the electrode engineers who had most to lose.

Established “Microbial fuel cells are a new idea.” Established Potter demonstrated microbial electricity generation in 1911 and Cohen stacked half-cells to over 35 V at 2 mA — about seventy milliwatts — in 1931. Speculative The idea is 115 years old and the power problem was legible in its first two decades, which is the relevant base rate for any claim that this decade's anode material changes things.

Established “Nocera's artificial leaf is ten times more efficient than photosynthesis.” Established The claim as made is real; the number readers infer from it is wrong by about an order of magnitude, because the comparison baseline is real-world plant efficiency, which is usually below one per cent — not photosynthesis's theoretical ceiling of 4.6 to 6.0 per cent. Established So “ten times” means single-digit per cent, not the 46 to 60 per cent a reader would get from the theoretical ceiling, which would be physically impossible for a single-junction device.

Established “Algae yield ten to twenty-three times more oil per hectare than oil palm, so algal fuel is close.” Established The yield range of 58,700 to 136,900 litres per hectare per year is a projection under ideal conditions, and the economics never closed at any yield: $59 to $186 per barrel against sub-$20 oil in the 1990s, $0.54 to $10.20 per kilogram of biomass in 2017, about $2.80 per litre of recovered oil, and ExxonMobil's exit in December 2022. Speculative Yield per hectare was never the binding constraint; harvesting dilute biomass from water, contamination at pond scale, and price were.

Established “The Aquatic Species Program proved algal fuel does not work.” Established It spent $25 million over eighteen years, screened about 3,000 strains, and was terminated in 1996 against sub-$20 oil; its durable scientific finding was a specific biological trade-off — rapid growth and high lipid accumulation are mutually exclusive under nitrogen starvation — not a general impossibility. Speculative Both readings in circulation misstate it: it identified one biological constraint and then ran out of money at a price point that no longer exists.

Speculative The minority position on that constraint deserves a hearing. If lipid accumulation can be decoupled from the starvation signal by direct pathway engineering — constitutive high-flux triacylglycerol synthesis in a nitrogen-replete cell — the trade-off dissolves. Handwave Thermodynamically this is doubtful, because the carbon budget is conserved and carbon diverted to lipid is carbon not available for growth; the realistic version is a shift in the trade-off curve rather than its elimination, and anyone claiming to have broken it should be asked for a closed carbon balance before anything else.

Established “Cellulosic ethanol is the next generation.” Established It was the next generation in 2007, at an estimated $2.65 per gallon against corn ethanol at a third to a half of that, and most plants were cancelled or closed in the early 2010s as the technical obstacles proved too difficult, and as of 2018 exactly one cellulosic ethanol plant remained operational. Speculative Anyone presenting cellulosic ethanol as forthcoming should be asked which plant.

Established “Biofuels are a marginal land use.” Established Up to forty per cent of the United States corn crop and roughly ten per cent of all world grain go to biofuel, and IEA figures for 2021 put ethanol production at twenty per cent of world sugar supplies and thirteen per cent of corn supplies — whatever else biofuels are, they are not marginal.

Established “Artificial photosynthesis is nearly here.” Established The highest reported laboratory prototype efficiency is 22.4 per cent, and the field's own summary is that no practical system has been demonstrated despite intense efforts since the 1990s and that the technology remains economically noncompetitive. Speculative High laboratory efficiency and zero deployment have coexisted for thirty years: the gap is not in the photochemistry.

Established “The scale-up economics paper proved that big bioreactors are impossible.” Established It did not: Humbird's analysis lands at $37 per kilogram of wet cell mass for fed-batch, $51 for perfusion, and $22 with $2-per-kilogram plant protein hydrolysate — expensive, not infinite. Established Its distinctive claim, and the sentence the field does not quote, is that bioreactors larger than 20 m³ have a lower maximum cell density because of carbon dioxide inhibition, with no route around it in a sparged vessel except sparging harder, possibly to the point of cell death. Speculative That is a named physical ceiling on vessel size, and it constrains biological energy proposals far more than the media-cost argument the paper is usually reduced to.

Speculative Finally, the exotic end, stated at full strength and then answered. The strongest version of “engineered electroactive organisms will beat photovoltaics plus electrolysis on an area basis” runs: biology self-assembles at ambient conditions, repairs itself, tolerates dilute and contaminated feedstocks, and could in principle be engineered so that cytochrome filament density, biofilm thickness and electron flux are design variables rather than wild-type accidents. Handwave Push it further and you get a photon-to-product organism with maintenance externalised to an engineered support system. Established What the mainstream says back is arithmetic: the areal record is 0.83 to 7.72 W/m² against roughly 200 W/m² for silicon at peak, practical algal conversion is about three per cent against about twenty, and the ceiling is not a materials problem but the cost of being alive. Speculative The honest terminal position is a declared tie on the framing and a loss on the numbers: the case for biology in molecules is strong and the case for biology in bulk electrons has not moved in a century, and the field's own best result — a coulombic efficiency over eighty-five per cent on a laboratory cathode in 2010, with no commercial plant sixteen years later — is the pattern in miniature.